Package jline.api.spn

Class Spn_structprops

java.lang.Object
jline.api.spn.Spn_structprops
public class Spn_structprops extends Object
Structural properties of a stochastic Petri net decided from its incidence matrix C = Post - Pre by linear programming, and its net-class membership.

Each LP property is decided per coordinate with ONE linear program, max sum(t) s.t. 0 <= t <= 1, t <= z, z >= 0 in the cone, whose optimum is unique with t = 1 exactly on the coordinates some cone vector covers; the property holds when all are covered:

  • bounded: y > 0, y'C <= 0 (boundedlevels: y(p) > 0)
  • conservative: y > 0, y'C = 0 (conservedlevels)
  • consistent: x > 0, C x = 0 (consistentevents)
  • repetitive: x > 0, C x >= 0 (repetitiveevents)
Net classes (each requires an ordinary net): state machine, marked graph, free choice, extended free choice, asymmetric choice.

Everything here is STRUCTURAL: inhibitor arcs, firing priorities and the precedence of immediate events are ignored; ignored lists those the net carries. Structural boundedness still implies boundedness of the GSPN.

References: T. Murata, Proc. IEEE 77(4) (1989), Secs. II-C and VII; M. Silva, E. Teruel, J.M. Colom, LNCS 1491 (1998) 309-373.

  • Method Details

    • spn_structprops

      public static Spn_structprops.SpnStructProps spn_structprops(NetworkStruct sn)
      Structural properties of a net.
      Parameters:
      sn - a NetworkStruct of a closed model holding Places and Transitions
      Returns:
      the structural verdicts